What is the highest grit sandpaper? I see from 2000-12000 grit

What is the highest grit sandpaper? I see from 2000-12000 grit paper. I went to a machine shop to see if he could mill down my heatsink and his machine would not be able to hold the heatsink and he recommended some 1200 grit sandpaper but on amazon i see stuff from 3M up to 2000 and other brands as high as 12000.

HF

so from a yahoo post it can go as high as 100k so what i was thinking was maybe use a 1000 grit paper to get my heatsink to be mostly even than use an 4k-20k for final touches. I can see that i would use too much paper trying to use a 8k+ paper the entire way. Probably would take forever and be expensive.

How fine is 4k grit paper anyways? Is that good enough or over kill or too rough?

I see a 3m product that has several types in it 1000, 1500, 2000, 2500, 3000. So i could slowly increase grit as i level it out. Where should i start and stop on grit?



The last thing is how do you know if their claimed grit is accurate? Are certain brands good and others are not? Or are all brands pretty much the same?


I bought these two off amazon and if i need to get a finner surface i'll look into the 8-12k ones i have seen

http://www.amazon.com/gp/product/B005JPGTNI/ref=oh_details_o02_s00_i00?ie=UTF8&psc=1

http://www.amazon.com/gp/product/B005RNGL9O/ref=oh_details_o08_s00_i00?ie=UTF8&psc=1

I also bought this one too so i could be good

http://www.riogrande.com/Product/3M-WetorDry-Tri-M-Ite-Polishing-Paper-8000-Grit/337306?Pos=10

If this works i'll let you guy's know. Read somewhere online that other people use this stuff on heatsinks.

Oh also the guy said use flat glass as a surface...he said something about glass being one of the smoothest and even surfaces that you can easily get your hands on. Since i don't really have a piece on milled still or anything lie that i'll give it a shot.


*fingers crossed*


*UPDATE: I went to advance to get one set of the 3m assorted and i have pretty much used all 5 sheets and it appears to be working except in two areas i still have no pressure....at all. The gap between the 3 areas must be huge! So i am glad i am buying 2 sets of amazon and a set of the 3k and a set of the 8k grit. I am thinking i might want to buy 2 more sets of the assorted since i am still not getting any sign of contact in those two areas. The assorted is the cheapest so that'll get ride of the bulk of the metal...at least i assume the the lower grit is more durable. I might just buy more actually....i kind want to try this on my R4 as well since the pressure on the R4 is kinda weak as well. Though i have to look at the readings. I thin the R4 is more lacking in pressure than unevenness. I have to look back at my other pressure paper to see my notes. If this works this may be the best way to get good cooling in a laptop....as long as it is not a pressure issue just an unevenness issue. I did also have a pressure issue but i grinded (SP?) down the mounts to make more pressure. This stuff is wet or dry paper. It says it lasts longer when wet but i am unsure how that really makes it last longer. I have been doing it dry. I might do it wetted (sp?) down with the 3k and 8k if the 8k is recommended wet.


Well i am using a microfiber cloth to clean up the sand paper to get some more use out of it. Not sure how the quality of the sand paper is and if it'll cause issues but i have a ton of copper to go to actually get a pressure reading so i am not too concerned since i'll touch it up with the new stuff when it comes. I am impatient what can i say.


Well i went through it again with the sand paper and i don't see much improvement since last time so i must have to remove a lot to get a reading :/ this may take awhile.

Also i lost a screw with taking heatsink off quickly in order to not mess up readings.....not sure where it went. Can't find it on desk or floor. I have a spare screw but i am afraid it got inside the laptop. Sigh may have to completely take it apart.

Well took it half apart and shook it all up and nothing fell out or sounded loose so i am hoping it isn't in there -_- No where to be found

 

making a new post to add more info.

So i tried working on it some more today and i think i might have made it worse. Also are CPU dies uneven anyways? I wonder if there is a point of where you can't actually do anything to make better contact.

Anyways i think i made it worse...i am not really sure. I have noticed i have pulled off far more than enough copper for it to be even without a doubt so i must not be applying even enough pressure or the mirror i am using isn't good enough. I am going to rip off a glass door off some furniture downstairs and give that a try when my new paper comes in. I want want to mess up the heat sink any further since i know my sandpaper i got at advance is past due.

This is really frustrating. I am also going to walk around the farm and see if i can acquire anything that will help the process. :/

I also have 1 spare heat sink so worse case i have a fall back...took me forever to get that one heat sink and none exist besides i one i barely found so my options of getting this to work is limited.

I could also try to de-solder (sp?) the copper block on the heat pipes and see if that machine shop can make me a new one :/

 

Hi there HoplesslyFaithful! I do a fair amount of car bodywork repair, so I know what it takes to get a surface flat..
I assume that you are trying to get the mating surface of your heatsink PERFECTLY flat so that it has the maximum surface contact area..
A good test to see if it is flat, is to very lightly spray the bottom of your heatsink with some cellulose spraypaint (I prefer matt black). Then SOAK your wet and dry paper (use 1200 for the test) with soapy water (prevents it from clogging up).
Put your paper on a flat surface (your mirror should be fine) and lightly rub your heatsink across it.

Keep checking the bottom of the heatsink. If there are any area's of the heatsink which are not flat, they will still be black (Paint not touching paper doesn't get removed).
Keep soaking the paper and rinsing the heatsink off.
Once all the black paint is gone, you know that every part of the heatsink is in contact with the paper, and must be relatively flat.
The only inconsistencies will be in your mirror and the surface of the chip you are trying to cool.
Heatsink paste is SPECIFICALLY designed to take care of this problem. It bridges any minor gaps and encourages better heat transfer.

Good luck!

 

Thanks for the reply. I think my problem was i was pushing too hard. This stuff is actually really aggressive. I thought i had a lot to remove but what was happening was my pressure was varying and in the back and forth swipes the pressure would shift. I am going to give it another shot in the near future and apply almost no pressure to take if slow and see how that works. Also I'll try it with water too since you talked about clogging. I think clogging might have also caused uneven sanding.

Also the gaps on laptop heatsinks are HUGE! With 4-8 mils of heatsprint I still don't get pressure readings on parts of the CPU. That is horrible pressure readings. I am also curious on how even the die actually is but i have no real way to test that and i have never found anything online about it

I did find this. I'll see if there is any major issues with the CPU in evenness.

 

Try holding either side of the heatsink, rather than the top.. Your fingers will act as a kind of crude pivot point.
The razor blade theory would work, but only if the blade is completely flat and true.. You can't GUARANTEE the razor blade to be true.
This reminds me of a conversation on how the first straight edge was invented.. What did they compare it against?

You talk about 4 to 8 millimeters of difference on heatsinks.. Thats disgusting. Unacceptable. Lego bricks are more accurate than that!

However, if you use the mirror and wet and dry paper method, you should get LESS than 1 millimeter difference across the surface (providing the mirror is true). The trick is in the paint.. Spray the bottom of the heatsink and let it dry. Just a quick burst of paint from 20cm away will do the trick. You don't need to completely change the colour of the bottom, just "tan" it.
Once the paint is dry, start rubbing the heatsink on the paper... High spots on the heatsink will have the paint almost instantly removed. You'll soon see where the imperfections are.

This method is used on luxury cars across the globe.. The paint on £700,000 Rolls Royce cars has to PERFECT!

 

yea i wasn't going to take it as fact...just wanted to see if there MIGHT be a large issue. If i don't see anything dramatic i won't care. I also wonder what the real difference is between a lapped IHS which is soldered and direct die contact. I wonder if there really is a difference if the CPU die is actually not smooth. If you can't get even pressure on a die than a soldered IHS might be better. I read somewhere between a soldered IHS and a direct die contact there wasn't a major difference....i could be wrong. I forget where that was. I only heard that once though so take that with a grain of salt.

 

WOOOOAHHH! Hold your horses!! The heatsink issue is something I can relate to because its basic engineering. I'm good with physics and electronics, but the ins and outs of socketed VS soldered? I'm just not that experienced with motherboards on PC's / laptops etc..

Soldered may actually pick up heat from the board itself and transfer heat to the chip in question.. There are so many variables that you're starting to talk about laboratory testing rather than home hacking.

I suppose it depends on what you want to achieve..
All i want is a working machine. I use my Xbox for gaming so my PC 's cruise along doing browsing, basic video editing, audio editing radio programming. In other words, any old crap with over 1GHz clock and more than 2GB ram will work for me!
You on the other hand seem to be testing he boundary's of what is possible.. A TRUE hacker!

I have a mobile PC (carputer but not in a car) here that I may be asking about soon.. I'd like to eliminate the fan and improve the heatsink etc.. You sound like you're heavily involved with cooling so I'll ask you first when I get to putting a thread together.

Let us know how it goes with your heatsink though.. I'm really interested in how it all turns out now!

 

I don't get that involved ^^ i was talking about a soldered IHS (integrated heat spreader) vs no IHS. Laptops have no IHS while desktops have an IHS. Normally IHS are soldered to the die but with IB they used a cheap TIM.

http://www.legitreviews.com/article/402/2/

this is the IHS on a dekstop and removed. I don't think you can actually remove it when it is soldered....which is what that article was trying to do. IB uses a TIM instead of solder which made it real crappy and allowed people to remove the IHS.

EDIT: I might give it a shot again tomorrow/today...whenever i wake up today

 

if you want it flat, you should be using Stuarts micrometer blue on a "AAA" granite or cast iron surface plate - and high precision machine shop will have these items, the thing is you absolutely do no use the surface plate as a grinding surface unless you intend on having the plate re-ground afterwards (this cost's £250-£1000+ for a AAA finish!!!!)

** another edit, micrometer blue is not a grinding paste, it is for measuring surface finish - you can pick up a 2nd hand surface place for £10-20 on ebay, and micrometer blue is £5 a tube..... you can combine the 2 to check the flatness of the finish you have - and yes, if your trying to make something flat by hand very, very gentle pressure is required, I'd recommend a fine grit diamond paste, and a piece of glass (one side of a sheet of glass is almost perfectly flat - again, very gentle pressure

Surface plate - Wikipedia, the free encyclopedia

in terms of surface finish, anything beyond 2-4k and your not sanding, your polishing

If you want something this small truly flat, a professional grinder shop might be able to help you with this, as you don't mill something if you want it perfectly flat, you would do it with a surface grinding machine - it's not really something you would do a such a small heatsink but this is the proper engineered approach. they can also then take the grinding wheel out, and fit a polishing mop, using the ridgidity of the machine and consistent distance/pressure applied to polish the surface without warping the surface finish (if you polish by hand it's impossible to apply the exact same amount of polishing pressure/time to every square mm of the surface, see the machine below, this is what you require (this or a horizontal milling machine fitted with a grinding wheel and then a polishing mop.

Surface grinding - Wikipedia, the free encyclopedia

you would do the CPU IHS (if it has one of course), and the heatsink to be sure. - if it's a laptop, then just the heatsink - lol don't polish or grind the surface of the silicone.....

Also regarding your IvyBridge you can replace that TIM with a similar liquid metal, I can source some if you want (*** edit, I can't post this accross a border, its requires a licence, so you'd have to source it locally), chemical compositions typically have a melt between 90(indigo extreme) and 134 to 170 °C I recommend the latter (know as base alloys) as they have no dimentional change after cooling, some of the compositions can shrink, some can expand a bit, some shrink then expand to the original size, it's a mine field, but in general "Base Alloy" is probably best (i'd have to check it's thermal conductivity and wetting ability but such a wide range of compositions is available you can usually find something to suit ie. high wetting, and high thermal conductivity)

one of the things about indigo extreme is it is probably one of the eutectics so all of it melts at the same time meaning it has no "molten" state where it is slightly lumpy

 

HAHA!

just seen your other thread, regarding indium, this part of the "base alloy" group, and sit's at about 156.6°C and has a thermal conductivity of 81.8w/mk, if your're getting 87 as you suggest, then you have a "Base Alloy" not pure Indium - which is an oxidising agent.... probably best not to use pure Indium..........

anyway, looks like your already having fun with this.

I'd recommend you try a few alloys with lower melting points, closer to say 70/80/90 °C - this way the alloy will start to flow as the CPU gets up to temperature, further "wetting" the contact area's and increasing thermal conductivity =) it won't last as long, but who cares, this is about performance, if we cared about longevity, we wouldn't be clocking to 4+ Ghz in tiny laptops

Have fun!